Skip to content

_tensor_pow

tensor_pow

Computes x^power element-wise along tensor.

This method can be used with Numpy data: 

n = np.array([[1, 4, 6], [2.3, 0.5, 0]])
b = fe.backend.tensor_pow(n, 3.2)  # [[1.0, 84.449, 309.089], [14.372, 0.109, 0]]
b = fe.backend.tensor_pow(n, 0.21)  # [[1.0, 1.338, 1.457], [1.191, 0.865, 0]]

This method can be used with TensorFlow tensors: 

t = tf.constant([[1, 4, 6], [2.3, 0.5, 0]])
b = fe.backend.tensor_pow(t, 3.2)  # [[1.0, 84.449, 309.089], [14.372, 0.109, 0]]
b = fe.backend.tensor_pow(t, 0.21)  # [[1.0, 1.338, 1.457], [1.191, 0.865, 0]]

This method can be used with PyTorch tensors: 

p = torch.tensor([[1, 4, 6], [2.3, 0.5, 0]])
b = fe.backend.tensor_pow(p, 3.2)  # [[1.0, 84.449, 309.089], [14.372, 0.109, 0]]
b = fe.backend.tensor_pow(p, 0.21)  # [[1.0, 1.338, 1.457], [1.191, 0.865, 0]]

Parameters:

Name Type Description Default
tensor Tensor

The input tensor.

 required
power Union[int, float]

The power to which to raise the elements in the tensor.

 required

Returns:

Type Description
Tensor

The tensor raised element-wise to the given power.

Raises:

Type Description
ValueError

If tensor is an unacceptable data type.
Source code in fastestimator/fastestimator/backend/_tensor_pow.py 
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
def tensor_pow(tensor: Tensor, power: Union[int, float]) -> Tensor:
    """Computes x^power element-wise along `tensor`.

    This method can be used with Numpy data:
    ```python
    n = np.array([[1, 4, 6], [2.3, 0.5, 0]])
    b = fe.backend.tensor_pow(n, 3.2)  # [[1.0, 84.449, 309.089], [14.372, 0.109, 0]]
    b = fe.backend.tensor_pow(n, 0.21)  # [[1.0, 1.338, 1.457], [1.191, 0.865, 0]]
    ```

    This method can be used with TensorFlow tensors:
    ```python
    t = tf.constant([[1, 4, 6], [2.3, 0.5, 0]])
    b = fe.backend.tensor_pow(t, 3.2)  # [[1.0, 84.449, 309.089], [14.372, 0.109, 0]]
    b = fe.backend.tensor_pow(t, 0.21)  # [[1.0, 1.338, 1.457], [1.191, 0.865, 0]]
    ```

    This method can be used with PyTorch tensors:
    ```python
    p = torch.tensor([[1, 4, 6], [2.3, 0.5, 0]])
    b = fe.backend.tensor_pow(p, 3.2)  # [[1.0, 84.449, 309.089], [14.372, 0.109, 0]]
    b = fe.backend.tensor_pow(p, 0.21)  # [[1.0, 1.338, 1.457], [1.191, 0.865, 0]]
    ```

    Args:
        tensor: The input tensor.
        power: The power to which to raise the elements in the `tensor`.

    Returns:
        The `tensor` raised element-wise to the given `power`.

    Raises:
        ValueError: If `tensor` is an unacceptable data type.
    """
    if tf.is_tensor(tensor):
        return tf.pow(tensor, power)
    elif isinstance(tensor, torch.Tensor):
        return torch.pow(tensor, power)
    elif isinstance(tensor, np.ndarray):
        return np.power(tensor, power)
    else:
        raise ValueError("Unrecognized tensor type {}".format(type(tensor)))